摘要
基于探地雷达回波信号进行处理以识别地下埋设的目标,始终是困扰雷达应用的关键,根据数据时间轴截断抑制直达波,利用宽相关处理进行信号滤波和典型数据自动提取,提高回波信号的信噪比.针对提取的典型道数据运用Welch功率谱处理,得到的特征数据归一化处理后作为径向基函数神经网络的输入,实现对地下埋设目标材质的自动识别与分类.在此基础上,分析了不同截断点对目标材质识别结果的影响.实测数据处理表明,本方法可以有效地实现对Fe、Al与土壤的识别和分类.
Basing on the echo signal to identify the underground targets is the key problem of me application of ground-penetrating radar. Data-time axial is used to reject and restrain the direct wave, and widehand correlation processing is used to filter to improve the signal to noise ratio （SNR） and extract the typical road data. The Welch power spectrum result of the typical road data is used as the input of radial basis function （RBF） network after the normalization processing, accomplishing the underground target automatic material identification and classification. Besides, it gives an effective analysis on the influence of the different cutting points to the results of target material recognition. The processing results of real data indicate that it can distinguish iron, aluminium and soil effectively.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2006年第1期98-102,共5页
Journal of Shanghai Jiaotong University
基金
上海市科技发展基金资助项目(015115038)
关键词
径向基函数神经网络
典型道数据提取
材质识别
探地雷达
radial basis function （RBF） neural network
typical road data extraction
material recognition
ground-penetrating radar
